[FEA]: Improve handling of thread and block indices

### Is this a duplicate?

- [x] I confirmed there appear to be no [duplicate issues](https://github.com/NVIDIA/cccl/issues) for this request and that I agree to the [Code of Conduct](CODE_OF_CONDUCT.md)

### Area

CUB

### Is your feature request related to a problem? Please describe.

We are currently quite inconsistent about our handling of threadidx and blockIdx, which leaves questions about overflow and integer promotions.

### Describe the solution you'd like

We should settle on a uniform approach, most likely a proper inline function that returns e.g `threadIdx.x + blockDim.x * blockIdx.x` in a safe maner

### Describe alternatives you've considered

_No response_

### Additional context

_No response_